Miller

Free plan LinuxMacSelf-hostedWindows

What Miller is

Miller is a command-line tool for querying and transforming data files. It handles formats such as CSV, TSV, JSON, JSON Lines, YAML, and DCF, alongside ASV, USV, PPRINT, Markdown, XTAB, DKVP, DKVPX, NIDX, and recutils. Users can add or remove fields, sort records, calculate aggregate statistics, format data for reading, and convert between formats. Miller works with named fields in key-value data, so transformations do not depend on counting column positions. It can also process input where records use different field names. Most operations work through records one at a time, which can allow suitable operations to handle files larger than available RAM. Miller is written in Go, has no runtime dependencies, and can be downloaded or compiled as a single binary. It runs on Linux, macOS, and Windows, with self-hosted deployment. Installation routes include package managers such as apt, yum, Homebrew, MacPorts, Chocolatey, Snap, and Conda, or building from source. The project repository identifies its license as two-clause BSD. Miller includes an MCP server in its executable for AI agents and MCP clients; it communicates over stdin/stdout using JSON-RPC and opens no network port. The documentation describes it as complementary to R, pandas, SQL databases, and Unix tools.

Who it suits

Miller suits people who work with data from the command line, including users in DevOps, data science, and biology or genomics. Its named-field operations and support for many formats are useful for workflows that reshape or inspect structured records.

What it does well

Works with many delimited, structured, and text data formats.
Transforms named fields without positional column counting.
Most operations process records one at a time.
Single binary with no runtime dependencies.
Includes an MCP server that opens no network port.

What to know first

Transformations use a code-based command-line interface.
Deployment is self-hosted.

Plans

Plan	Billed and limits	Price
Miller	No paid plans or usage limits stated	Free

As miller.readthedocs.io lists them, read 8 Oct 2026. The maker's own page has the final word.

Getting started

Visit the Miller documentation website.
Choose an installation route such as a package manager or building from source.
Install Miller on Linux, macOS, or Windows, or compile it as a single binary.
Use Miller from the command line to query or transform data files.

Verdict

Choose Miller if you want a free, open-source command-line tool for querying, reshaping, and converting structured data across many formats. Its record-at-a-time processing and named-field approach suit workflows that reshape or inspect structured records.

Questions

How much does Miller cost?

Miller is free. Its listed plan is free, and no paid plans or usage limits are stated.

Is Miller open source?

Yes. The project repository identifies its license as two-clause BSD.

Which platforms does Miller support?

Miller runs on Linux, macOS, and Windows, and supports self-hosted deployment.

Which data formats can Miller process?

Supported formats include CSV, TSV, ASV, USV, JSON, JSON Lines, YAML, PPRINT, Markdown, XTAB, DKVP, DKVPX, NIDX, DCF, and recutils.

How can Miller be installed?

Installation options include apt, yum, Homebrew, MacPorts, Chocolatey, Snap, and Conda, as well as building from source.

Does Miller work with AI agents?

Its executable includes an MCP server that lets AI agents and MCP clients discover and use Miller tools. The server uses JSON-RPC over stdin/stdout and opens no network port.

Facts from Miller's own pages

AI support: Miller includes an MCP server in its executable so AI agents and MCP clients can discover and use Miller tools. · miller.readthedocs.io, 8 Oct 2026
Data operations: It can add and remove fields, sort, aggregate statistics, pretty-print data, and convert between formats. · github.com, 8 Oct 2026
Distribution: Installation options include package managers such as apt, yum, Homebrew, MacPorts, Chocolatey, Snap, and Conda, as well as building from source. · miller.readthedocs.io, 8 Oct 2026
Integrations: The documentation describes Miller as complementary to R, pandas, SQL databases, and the Unix toolkit. · miller.readthedocs.io, 8 Oct 2026
Intended users: The author describes use across disciplines including DevOps, data science, and biology/genomics, while noting that many users focus on CSV. · miller.readthedocs.io, 8 Oct 2026
License: The project repository identifies Miller's license as two-clause BSD. · github.com, 8 Oct 2026
MCP transport: The MCP server uses JSON-RPC over stdin/stdout and opens no network port. · miller.readthedocs.io, 8 Oct 2026
Named fields: Miller works with named fields in key-value data without requiring users to count positional column indices. · miller.readthedocs.io, 8 Oct 2026
Platforms: The installation page lists binaries or installation methods for Linux, macOS, and Windows. · miller.readthedocs.io, 8 Oct 2026
Purpose: Miller is a command-line tool for querying, shaping, and reformatting data files in formats including CSV, TSV, JSON, JSON Lines, YAML, and DCF. · miller.readthedocs.io, 8 Oct 2026
Record schemas: Miller can process interleaved records with different field names. · miller.readthedocs.io, 8 Oct 2026
Runtime: Miller is written in Go and has no runtime dependencies; it can be downloaded or compiled as a single binary. · miller.readthedocs.io, 8 Oct 2026
Streaming: Most operations process one record at a time, allowing Miller to work with files larger than available RAM when the operation permits. · miller.readthedocs.io, 8 Oct 2026
Support: The project directs users to GitHub Discussions for general questions, Stack Overflow for tagged questions, and GitHub Issues for bug reports and feature requests. · miller.readthedocs.io, 8 Oct 2026

Where Miller stands on Freedom251

Data Transformation Tools#18 of 37